DESCRIPTION: Humans develop tolerance to caffeine after a few days of caffeine administration. Without a sufficient withdrawal period, this tolerance may mask the true physiological effects of the drug. There is some evidence that a withdrawal period of four days is sufficient to permit resensitization to the physiological effects of caffeine. However, it is not known if the full four days are required, or if a shorter time period is sufficient. The purpose of the proposed study is to determine the withdrawal period necessary to obtain a physiologically caffeine naive response by moderate caffeine consumers. Fifteen males and fifteen females of college age will be recruited as subjects. These subjects will be healthy, non- tobacco users who consume caffeine in moderate amounts (200-300 mg/day). Each subject will complete eight double-blind counterbalanced trials, separated by at least a week. The subjects will consume either caffeine (5 mg/kg body weight) or a placebo (dextrose) in gelatin capsules with 200 ml water at time 0. Each of these conditions will be performed following no withdrawal and the pre-trial caffeine withdrawal periods of 24 hours, 48 hours, and 4 days. Diet composition will be controlled for the 4 days preceding each trial, and the subjects will be instructed to abstain from exercise on the day before each trial. All trials for a given subject will be conducted at the same time of the morning following an overnight fast. A catheter will be inserted into the antecubital vein at minute -60. Blood samples will be taken at minutes -30, 0, 20, 40, 60, 80, and 120 for later analysis of serum glucose, serum free-fatty acids, plasma epinephrine and norepinephrine. Measurements of blood pressure, heart rate, and respiratory rate will be taken at minute -30, and every 10 minutes from minutes 0 to 120. Statistical significance will be determined using three way analyses of variance with repeated measures with drug, abstinence period, and time as factors. When a significant interaction occurs, the location of pairwise differences will be determined using a Tukey post hoc analysis (p<.05).